1. Field of the Invention
The present invention relates to a method of converting color data, and more particularly to a method of converting color data when the gamut of a color input/output device for input and that of a color input/output device for output are different.
Furthermore, the present invention relates to a method of modifying the registration data in a color conversion table which is referred to when the registration data are converted to a different expression system of color signals.
2. Description of the Related Art
Along with the recent advent of high-performance personal computers and low-cost color printers, opportunities to print color pictures have remarkably increased. When color pictures are printed, particularly the coincidence in appearance between color pictures displayed on a CRT display connected to a personal computer and those printed by a color printer is required.
There are a wide variety of expression methods of color signals. For example, printers and scanners express color using three primary colors, such as CMY values, RGB values, etc. Taking the output color of a printer in which the minimum signal value is 0 and the maximum signal value is 255, as an example, when color is expressed using CMY, all output colors are expressed by the combinations of CMY values ranging from 0 to 255. In this case, this method has a characteristic that even if data to be printed are prepared using the same CMY values, printed (reproduced) color varies with the kind of printer. For this reason, the CMY values are a way of expressing color, dependent on individual devices. A scope in which the colors of this individual device can be reproduced, is called a gamut.
Besides this expression method of color signals dependent on devices, there is an expression method of absolute colors independent of devices. This is a method of expressing colors using coordinate values, such as L*a*b*, XYZ, etc. Since colors expressed by this method are clearly defined, colors can be expressed independently of individual devices. Color expression methods such as L*a*b, XYZ, etc. are used as an intermediate color expression method when colors are converted between the color spaces of different devices. That is, this method is used in such a case where, for example, RGB values used in a first device are once converted to, for example, L*a*b* values, and then the converted L*a*b* value are converted back to RGB values again. If color signals are accurately converted by these expression methods, inputted colors (in the case of scanners, etc.) and outputted colors (display on display monitors or printing by printers) between different devices can be matched with each other.
In this way, in order to match color outputs between devices, color data have to be converted using a variety of expression methods of color data. For a method of converting color data, there is a method of using the correspondence of color in each color space, that is, a color conversion table in which paired data of first color data in a first device and paired data of second color data in a second device are stored. Generally speaking, in a conversion method of using a color conversion table, when a color registered in the color conversion table is converted, the color can be converted by using the registered correspondence as it is. Therefore, the color conversion is processed at a remarkably high speed. However, since data volume to be registered in the table becomes enormous, sometimes all the color correspondences are not registered in the color conversion table. In this case, when a color which is not registered in the color table is converted, colors in the neighborhood of the color to be converted are selected, by performing an interpolation operation process for color signals in the neighborhood, the color data to be converted is calculated, and the color conversion is performed based on this calculation result.
However, in the conversion of a color which is not registered in a color conversion table, errors occur in the interpolation operation process.
Although in preparing a color conversion table, operations to establish each color correspondence are performed, in these operations too, errors occur. For this reason, even when color conversion is performed using color data registered in the color table, errors often occur as a result of color conversion.
First, a conventional color data conversion method is described.
An example of conventional color data conversion methods and the used apparatuses is disclosed in the Japanese Laid-open Patent Publication No. 60-105376. In this example, first color data included in a first gamut dependent on a first color input/output device are converted toward an achromatic color with the same lightness as the first color data, and second color data included in a gamut dependent on a second color input/output device are calculated.
FIG. 1 shows the L*b* cross section of a L*a*b* space. A field where b* is positive is largely the cross section of a yellow field, and an field where b* is negative is largely the cross section of a blue field. L* indicates lightness, and as a* and b* values increase, the saturation also increase. A field enclosed by straight lines 3701 shown in FIG. 1 is a gamut in the L*a*b* space of a display using a CRT display being a general-purpose color input/output device. A field enclosed by straight lines 3702 is a gamut of a printer being a general-purpose color input/output device using inks of cyan, magenta and yellow (CMY).
When first color data 3703 being yellow with high saturation which can be displayed on the CRT display are printed, this color cannot be printed because of a characteristic problem on color available from a printer. Therefore, the color data 3703 has to be converted in such a way that they can be printed. In a conventional method, the color data 3703 are converted using an algorithm toward third color data 3704 of an achromatic color with the same lightness value as the color data 3703 so as to be included in a gamut 3702, and the second color data 3705 are calculated. In the same way, the first color data 3706 being blue with high saturation are converted toward third color data 3707 of an achromatic color with the same lightness value as the lightness value of the color data 3706 so as to be included in a gamut 3702, and the second color data 3708 are calculated.
Since a printer can print color data 3705 and 3708, the printer prints the color data 3705 instead of the color data 3703, and the color data 3708 instead of the color data 3706. However, as clearly shown in FIG. 1, the color data 3705 are obtained with fairly low saturation compared with the color data 3703, and the color data 3708 are also obtained with fairly low saturation compared with the color data 3706. Using the same algorithm as described above, second color data 3711 are obtained from first color data 3709. Although before conversion the color data 3703 has higher saturation than the color data 3709, after conversion the color data 3711 corresponding to the color data 3709 has far higher saturation than the color data 3705 corresponding to the color data 3703, and the order in saturation degree is reversed between before and after conversion.
In a technology disclosed in the Japanese Laid-open Patent Publication No. 4-101566, first color data included in a first gamut dependent on a first color input/output device are converted toward an achromatic color of different lightness for each hue angle, and second color data included in a gamut dependent on a second color input/output device are calculated.
Fields and straight lines 3701 and 3702 shown in FIG. 2 are the same as those shown in FIG. 1. Since first color data 3721 being yellow with high saturation which can be displayed on a CRT display, cannot be printed by a printer, the color data 3721 are converted using an algorithm toward third color data 3722 of an achromatic color with high lightness so as to be included in a gamut 3702, and second color data 3723 are calculated. In the same way, color data 3724 being blue with high saturation are converted toward color data 3725 of an achromatic color with low lightness so as to be included in the gamut 3702, and second color data 3726 are calculated. Then, the color data 3723 and 3726 are printed by a printer. In this conventional method, the color data 3723 are converted with fairly low saturation compared with the color data 3721, and the color data 3726 are converted with a fairly low saturation compared with the color data 3724.
Using the same algorithm as described above, second color data 3728 are calculated from first color data 3727. Although before conversion the color data 3721 has a higher saturation than the color data 3727, after conversion the color data 3728 corresponding to the color data 3727 has far higher saturation than the color data 3723 corresponding to the color data 3721, and here again the order in saturation degree is reversed between before and after conversion.
In a thesis by P. Laihanen, "Colour Reproduction Theory based on the Principles of Colour Science", a technology is described in which all the first color data included in a first gamut dependent on a first color input/output device are converted toward an achromatic color with a mean lightness value between the highest lightness value and the lowest lightness value of a second gamut, and second color data are calculated.
Fields and straight lines 3701 and 3702 shown in FIG. 3 are the same as those shown in FIG. 1. Since first color data 3741 being yellow with high saturation which can be displayed on a CRT display, cannot be printed by a printer, the color data 3741 are converted using an algorithm toward third color data 3742 of an achromatic color with a mean lightness value between the highest lightness value and the lowest lightness value of a second gamut 3702 so as to be included in the gamut 3702, and second color data 3743 are calculated. In the same way, color data 3744 being blue with high saturation are converted toward third color data 3742 so as to be included in the gamut 3702, and second color data 3745 are calculated. Then, the color data 3743 and 3745 are printed by a printer. In this conventional method, as clearly shown in FIG. 3, the color data 3743 and 3745 are converted with fairly low saturation compared with the color data 3741 and 3744.
Using the same algorithm as described above, second color data 3747 are calculated from first color data 3746. Although before conversion the color data 3741 has higher saturation than the color data 3746, after conversion the color data 3747 corresponding to the color data 3746 has higher saturation than the color data 3743 corresponding to the color data 3741, and here the order in saturation degree is reversed between before and after conversion.
The conventional color data conversion methods have the following problems.
FIG. 4 shows an a*b* plane corresponding to a specific L* value in L*a*b* values. A field enclosed by straight lines 3801 shown in FIG. 4 is a gamut in the L*a*b* space of a CRT display being a general-purpose color input/output device. A field encircled by straight lines 3802 is the gamut of a printer being a general-purpose color input/output device using inks of cyan, magenta and yellow (CMY).
When first color data 3803 being yellow with high saturation which can be displayed on a CRT display are printed by a printer, since this color cannot be printed by a printer, the color signal has to be converted so as to be printed by a printer. In the above three conventional methods, the color data 3803 are converted toward third color data 3804 of an achromatic color with the same lightness value as that of the color data 3803, a lightness value calculated dependent on the hue angle value of color data 3803 or a mean lightness value between the highest lightness value and the lowest lightness value of a second gamut (a field enclosed by the straight lines 3802) so as to be included in the gamut 3802, and second gamut 3805 are calculated. In the same way, first color data 3806 being blue with high saturation are converted toward the third color data 3804 of an achromatic color with the same lightness value as that of the color data 3806, a lightness value calculated dependent on the hue angle value of color data 3803 or a mean lightness value between the highest lightness value and the lowest lightness value of a second gamut (a field enclosed by the straight lines 3802) so as to be included in the gamut 3802, and second color data 3807 are calculated. First color data 3808 being magenta with high saturation are converted toward the third color data 3804 of an achromatic color with the same lightness value as that of the color data 3808, a lightness value calculated dependent on the hue angle value of color data 3803 or a mean lightness value between the highest lightness value and the lowest lightness value of the second gamut 3802 so as to be included in the gamut 3802, and second color data 3809 are calculated.
However, in this conventional method, because of the non-linearity of equi-hue lines on a CRT display in an L*a*b* space, the color data 3806 of blue with high saturation are compressed into the color data of purple, and as a result, the color displayed in blue with high saturation on the CRT display changes to purple when the color is printed by a printer.
First color data 3808 of magenta with high saturation are converted to second color data 3809 of purple with low saturation for the reason that the color data 3808 greatly differs in hue angle value from the color data 3810 of magenta with high saturation of the printer. The problem that since the color data of the primary and secondary colors of a CRT display and the color data of the primary and secondary colors of a printer greatly differ in hue angle value, the saturation is significantly lost when the color data of the CRT display are converted by the printer, which occurs in the cases of red, yellow, green, cyan, etc. with high saturation in addition to magenta with high saturation.
For the above problem that blue changes to purple, one solution is disclosed in the Japanese Laid-open Patent Publication No. 6-233129. In this conventional method, the problem is solved by rotating and shifting colors in both blue and purple fields around an achromatic color axis along the distribution curve of the blue and purple fields in an XY chromaticity diagram.
FIG. 5 shows the details of this conventional method. FIG. 5 shows an a*b* plane at a specific L* value. 3821 and 3822 shown in FIG. 5 are the gamuts of a printer and a CRT display, respectively. A vertically shaded field 3823 is a field judged to be purple and a horizontally shaded field 3824 is a field judged to be blue, when seen by the human eye.
Here, a case where first color data 3825 displayed on a CRT display by this conventional method are printed by a printer is described. Since the color data 3825 cannot be printed by a printer, the color data are compressed toward third color data 3826 on an achromatic color axis. However, if the color data 3825 are compressed toward color data 3826 without taking into consideration the distortion of equi-hue lines, the color data 3825 of blue are compressed into color data 3827 of purple. For this reason, before the saturation of the color data 3825 is converted, the color data 3825 is rotated around the achromatic color axis and are shifted to color data 3828. Thus, the color data 3825 of blue are converted to color data 3829 of blue. However, in this method, since the fact that the distortion of equi-hue lines differs depending on the lightness is not taken into consideration, colors which are converted with the hue accurately preserved are limited to color data with a certain lightness value.
FIG. 6 shows the states of the distortion of equi-hue lines in a purple field at each lightness. FIG. 6 shows the gamuts of both a CRT display and a printer in the cases where the lightness value is 25, 50 and 75, and (a), (c) and (e), and (b), (d) and (f) are the gamuts of the CRT display and the printer, respectively. In each gamut, a vertically shaded area 3841 indicates a purple field and a horizontally shaded area 3842 indicates a blue field. Broken lines 3843 are equi-hue lines connecting colors with a certain hue. It is shown in FIG. 6 that the equi-hue line 3843 of the CRT display and the equi-hue line 3843 of the printer differ in angle and shape for each lightness.
Therefore, in this conventional method, only colors with a certain lightness can be converted in such a way that the hues can be matched between the CRT display and the printer. For this reason, pictures displayed on a CRT display cannot be faithfully reproduced by a printer.
Next, a conventional method of generating a color conversion table according to a variety of color data conversion methods, and performing a color conversion is described.
In order to prevent errors from occurring in color conversion using a color conversion table, it can be considered that data registered in the table are adjusted, for which there is a method disclosed in the Japanese Laid-open Patent Publication No. 8-287226. In this method, an L*a*b* color space is reconfigured as a grid-shaped (particularly three-dimensional grid-shaped) space, and CMY values corresponding to L*a*b* values distributed on each grid point are registered in the color table. The L*a*b* values of measurement data which are surrounded with grids and exist inside the grids (that is, do not exist on the grid points when CMY values registered in the color table are calculated), is calculated in such a way that both actual CMY values corresponding to the L*a*b* values on the point and the squared error of CMY values calculated by interpolation operation may become a minimum. FIG. 7 is a flowchart summarizing the above method. First, an L*a*b* space is reconfigured in a grid shape, and L*a*b* values are set at each grid point. Then, a color is actually outputted by a printer, etc., the CMY values are calculated, and simultaneously corresponding L*a*b* values are calculated. This process is performed for a plurality of grid points, and CMY values registered at the grid point are determined based on the relation of measurement data at a grid point by making a printer output actual colors.
In the conventional method described above, particularly colors included in both yellow and blue fields are converted with the saturation lost. For this reason, when pictures including yellow with high saturation displayed on a display monitor (pictures in which a banana, lemon, etc. are photographed by a digital camera and are prepared, pictures in which vivid printed matters are inputted by a color image scanner, etc.) are printed by a printer, a problem occurs that the pictures become unnatural, since the colors are printed with the saturation of yellow greatly lost. In the same way, when pictures including blue with high saturation (sea, blue sky, etc.) are printed by a printer, a problem occurs that the pictures become unnatural, since the colors are printed with the saturation of blue greatly lost.
In the above conventional method, the order in saturation of two color data is reversed between before and after conversion. For this reason, when color pictures are printed, the gradation is reversed, and a problem occurs that the pictures become abnormal and unnatural.
In the above conventional method, when blue with high saturation displayed on a CRT display is printed by devices with a narrower gamut, such as printers, etc., the distortion of equi-hue lines in a blue field is not taken into consideration, and a problem occurs that the blue changes to purple.
When the primary and secondary colors of a first color picture device and the primary and secondary colors of a second color picture device greatly differ in hue angle value, such as when pictures displayed on a CRT display are printed by a printer, a problem occurs that colors in the neighborhood of the primary and secondary color of the first color picture device are printed with the saturation greatly lost.
Alternatively, in the conventional method, when blue with high saturation displayed on a CRT display is printed by a device with a narrower gamut, such as a printer, etc., colors in both blue and purple fields are rotated and shifted around an achromatic color axis along the curved shape of both blue and purple fields in an xy achromaticity diagram, the saturation is then converted; and the colors are printed by a printer. However, a problem occurs that the distortion of equi-hue lines varying with the lightness is not taken into consideration, and as a result, out of blue displayed on the CRT display only a part of the blue with a certain lightness value can be printed with the hue preserved, and most of the blue changes to purple.
Furthermore, in the conventional method, although the errors of a plurality of measurement data can be minimized, it is difficult to guarantee the accuracy of a specific color or to guarantee the conversion results of a specific color.
Specifically, for example, the following requirements cannot be met.
(1) It cannot be guaranteed that CMY values/CMYK values for the L*a*b* values of the color of paper is perfectly made 0. PA0 (2) The color of characters overlaid on a picture inputted by a scanner, etc. becomes unnatural.
In a color conversion table for printing, CMY values and CMYK values corresponding to the color of L*a*b* values distributed in a grid shape are registered. For grid points, values of integers are selected. However, generally speaking, the L*a*b* values of the color of paper, such as "90.5, 1.2,-4.2" do not coincide with grid points, and are not integers. Therefore, when color conversion is performed for the color of paper, by performing an interpolation operation based on data registered in the grid points around a measured L*a*b* value, CMY values for "90.5, 1.2,-4.2" have to be calculated.
Since this color is originally the color of paper, when the color is printed, the ink volume has to be perfectly 0, that is, all values of CMY(K) have to be 0 (it must be a state where nothing is printed). However, since the L*a*b* values are not integers, calculation errors occur in the interpolation operation. For this reason, all the CMY values always become 0. Thus, in this case, a problem occurs that colors are printed on the paper background (ink is deposited on the paper background).
Generally speaking, there are errors in the result of color conversion, and it is often the case that a difference in colors cannot be recognized, if printed colors are somewhat different from colors to be printed. However, since this is a part where something should be printed, the difference in color is not conspicuous only, and nothing more. When even a little volume of ink is deposited on a piece of paper which should be on the paper background of a printing medium and on which there should be no ink, a human eye recognizes the change as a great difference, even if the change volume of color is a little. Therefore, it is desirable to convert in such a way that an error may not occur in an interpolation operation.
In a color conversion table for converting pictures inputted by a scanner (for example, RGB values) to L*a*b* values, L*a*b* values corresponding to RGB values distributed in a grid shape are registered.
Generally speaking, scanners are often designed to have a contingency in output signals. That is, the scanners are designed in such a way that signals representing colors read from an input picture may not take extreme values. For this reason, in such a case, the scanners are designed so that neither the maximum value of 255 nor the minimum value of 0 may be outputted as an R value, G value or B value. For this reason, in such a case, in a color conversion table prepared based on the actually measured values of color, for example, L*a*b* values corresponding to a grid point of RGB=(255, 255, 255) cannot be directly calculated. This is because the data of RGB=(255, 255, 255) are not outputted from a scanner.
Thus, in order to calculate L*a*b* values corresponding to RGB=(255, 255, 255), anticipated and prepared L*a*b* values are stored by performing some process, such as extrapolation, etc.
The L*a*b* values which are anticipated and prepared by performing some process, such as extrapolation, etc. described above are not always integers. For example, a value of a number with fractions such as (98.5, 1.2, 7.2) is registered. As described earlier, in a scanner designed in such a way that the maximum value of 255 may not be outputted, since the data of RGB=(255, 255, 255) are not outputted, even if a picture is scanned, there is no problem in color conversion for the scanned picture, even if any kind of L*a*b* values are registered at this grid point.
However, characters, etc. are often overlaid on a picture inputted by a scanner. For example, when a white character is overlaid on a picture, a picture element corresponding to the character takes a value, such as RGB=(255, 255, 255). Here, it is generally understood that the value of RGB=(255, 255, 255) means white in a display monitor.
When by using a color conversion table registering a value of (98.5, 1.2, 7.2) for RGB=(255, 255, 255), a picture on which characters are overlaid as described above is converted, the part of characters becomes L*a*b*=(98.5, 1.2, 7.2). If L*a*b* values of (98.5, 1.2, 7.2) are assumed not to represent white, a problem occurs that although the part is a portion where originally white is intended and characters are overlaid, by color conversion the characters become colored.
Although as concrete problems, the color of paper in (1) and overlaid white characters in (2) have been so far described, regarding (1), the problems occur for a variety of colors, for example, that yellow with high purity reproduced using Y (yellow) ink is mixed with ink of another color. Problems on the color of overlaid characters in (2) are not limited to white, but also occur in a variety of colors, such as red RGB=(255, 0, 0), black RGB=(0, 0, 0), etc.
The present invention has been made in order to solve the problems described above, and it is an object of the present invention to provide a method of converting color which can be displayed on a first color picture device, particularly both yellow and blue with high saturation, to color data which can be displayed on a second color picture device without degrading saturation.
Furthermore, it is another object of the present invention to provide a method of converting two pieces of color data which can be displayed on a first color picture device, to color data which can be displayed on a second color picture device without reversing the order in saturation between the two pieces of color data.
Furthermore, it is another object of the present invention to provide a method of converting pictures displayed on a first color picture device, such as a CRT display, etc., to color data for a second color picture device without changing all the hue values of the first color data in order to faithfully reproduce the pictures on the second color device, such as a printer, etc.
Furthermore, it is another object of the present invention to provide a method of converting pictures displayed on a first color picture device, such as a CRT display, etc., to the color data of the primary and secondary colors with high saturation of a second color picture device without the loss of the vividness of the primary and secondary colors with high saturation of the first color picture device in order to reproduce the pictures on the second color picture device, such as a printer, etc.
Furthermore, it is another object of the present invention to implement a method of converting the registration data of a color conversion table such that errors based on the interpolation operation can be removed.
According to the first aspect of the present invention, in a color conversion method of generating second color data included in the second gamut of a second color picture device from the first color data included in the first gamut of a first color picture device, the color data conversion method of the present invention comprises the steps of calculating third color data which are included in the second gamut and correspond to the saturation value of the first color data, and calculating the second color data to convert the first color data toward third color data included in the second gamut.
Then, particularly in the step of calculating the third color data, the third color data are calculated in such a way that the data may be included in the second gamut, and the greater is the saturation value of the first color data, the greater may become the saturation value of the third color data.
Furthermore, in the step of calculating the third color data, fourth color data with the highest saturation of the color data included in a first gamut are calculated in the same hue as the first color data, fifth color data with the highest saturation of the color data included in a second gamut are calculated in the same hue as the first color data, and third color data are calculated in such a way that the third color data can be included in the second gamut, the greater is the saturation value of the first color data, the greater may become the saturation value of the third color data, and the saturation value of the third color data may become nearly equal to the saturation value of the fifth color data, when the saturation value of the first color data is nearly equal to the saturation value of the fourth color data.
Furthermore, in the step of calculating the third color data, fifth color data with the highest saturation of the color data included in a second gamut are calculated in the same hue as the first color data, a first difference in lightness value between the first color data and the fifth color data, and third color data are calculated in such a way that the third color data may be included in the second gamut and the saturation value of the third color data can be determined by two factors, one being a factor in which the greater is the saturation value of the first color data, the greater becomes the saturation value of the third color data, and the other being a factor in which the greater is the first lightness difference value, the smaller becomes the saturation value of the third color data.
Furthermore, in the step of calculating the third color data, the third color data are calculated on a first straight line connecting sixth color data being an achromatic color and the fifth color data.
Furthermore, in the step of calculating the third color data, the sixth color data are for an achromatic color with lightness value nearly equal to that of the fifth color data, and in the step of calculating the third color data, the sixth color data are for an achromatic color with lightness value nearly equal to a mean value between the maximum lightness value and the minimum lightness value of the second gamut.
Furthermore, in the step of calculating the second color data, a second straight line is connected between the first color data and the third color data, and color data positioned at the intersecting point of the second straight line and the outermost boundary of the second gamut are calculated as second color data, the first color data and the third color data are connected by the second straight line, seventh color data positioned at the cross point of the second straight line and the outermost boundary of the first gamut are calculated, eighth color data positioned at the cross point of the second straight line and the outermost boundary of the second gamut are calculated, and color data obtained by dividing the part of a line on the second straight line surrounded with the third color data and the eighth color data based on the ratio of distance between the third color data and the seventh color data to the distance between the third color data and the first color data are calculated.
According to the second aspect of the present invention, in a color data conversion method of generating second color data included in the second gamut of a second color picture device, a color data conversion method comprises the steps of calculating a shift amount value from the hue angle value, lightness value and saturation value of first color data, and converting the hue angle value of the first color data based on the shift amount value.
In the step of calculating the shift amount value, a multi-dimensional table is used to calculate the shift amount value.
Alternatively, when the first color data are color data representing blue or purple of the first color picture device, the step of calculating the shift amount value is a process of calculating the shift amount value to reduce the hue angle value.
Furthermore, when the first color data are color data representing colors in the neighborhood of the primary and secondary colors of the first color picture device, the step of calculating the shift amount value is a step of calculating a shift amount value to convert the hue angle value of first color data to that of color data representing color in the neighborhood of the primary and secondary colors of the second color picture device.
Alternatively, the step of converting the hue angle value of the first color data based on the shift amount value is a step of rotating the first color data around the color data of an achromatic color with the same lightness as the first color data based on the shift amount value, and converting the hue angle value of the first color data.
In a color conversion system for generating the second value of a color space which is included in the second gamut of a second color picture device and is dependent on the second color picture device, from the first value of a color space which is included in the first gamut of a first color picture device and is dependent on the first color picture device, the color conversion system of the present invention comprises a unit for converting the value of the color space dependent on the first color picture device to the third value of the color space independent of devices, a unit for calculating a shift amount value from the hue angle value, lightness value and saturation value of the third value of the color space, a unit for converting the hue angle value of the third value of the color space based on the shift amount value, and a unit for converting the third value of the color space of which the hue angle value is converted, to the second value of a color space dependent on the second color picture device.
The modification method of registration data registered in the color conversion table of the present invention in which the correspondences between colors before and after conversion to be referred to at the time of conversion are stored, comprises the steps of referring to the correspondences of colors to be guaranteed, selecting data of which the registration value is modified, and modifying the registration data so as to guarantee the referred correspondence.